You can use this analysis object to derive univariate collectives from an existing Rainflow matrix or Markov matrix. These one-dimensional results can be plotted and compared more easily.

Procedure

The following collectives can be derived from the matrices:

The "positive" and "negative" results can be counted separately for each collective. For the peak value count, these are, e.g. the peak and trough values. For the level crossing count, these are the upward and downward crossings.

Note:   If you use the Rainflow matrix, this must be present in  the "from"-"to" format.

Frequency

The count can determine absolute, relative and cumulated frequencies. The sum of the absolute frequencies must be equal to the number of events counted. This can be lower than the number of values in the source data set! For relative and percentage frequencies, normalization is to one or one hundred, i.e. the sum of the frequencies results in one or one hundred. If the signal contains void values, this is accounted for when normalization takes place. For the cumulated frequencies, the individual frequencies are summed up from the top, in the case of positive events, and from the bottom for negative events. In the case of a peak value count, the value for a class specifies the number of peaks in this class and of all of the higher classes. For the trough values, accordingly, it specifies the number of troughs in this class and all of the lower classes. No cumulated frequencies may be selected for the level crossings count, since its result is already cumulated. This results from the fact that larger ranges cross several class limits. For the range count and range pair count, the first range has by definition the length zero and thus also the frequency zero. This definition has the advantage that the number of frequencies in this procedure is also equal to the number of classes. Analogous to this, in the case of level crossing count, the lowest class limit is also included in the collective, although this is never exceeded, since the values below this limit are ignored. As a result, the analysis object provides a signal whose Y component contains the frequencies counted. The contents of the X component depends on the count procedure:

* The X component of the result is calculated from the class mid-points, which reside in the X component of the counted matrix. This is only possible, however, if the count matrix is based on an equidistant class division.

References

[1] de Jonge, J.B. (1980). Counting Methods for the Analysis of Load Time Histories. NLR Memorandum SB-80-106 U

ASTM E1049-85(2017). Standard Practices for Cycle Counting in Fatigue Analysis. STM International, West Conshohocken, PA, 2017

FPScript Functions Used

LevelCrossingCount

RangeCount

PeakCount

See Also

Counting Procedures Option

Analysis Objects

Count Matrix Analysis Object

Count Analysis Object

Compound Count